In recent years, thrombolytic therapy with fibrinolytic agents, such as Streptokinase (SK), tissue plasminogen activator (TPA) or urokinase (UK) has revolutionized the clinical management of diverse circulatory diseases e.g., deep-vein thrombosis, pulmonary embolism and myocardial infarction. These agents exert their fibrinolytic effects through activation of plasminogen (PG) in the circulation by cleavage of the scissile peptide bond between residues 561 and 562 in PG. As a result, the inactive zymogen is transformed to its active form, the serine protease, plasmin (PN), which then acts on fibrin to degrade the latter into soluble degradation products. It may be mentioned here that PN, by itself, is incapable of activating PG to PN; this reaction is catalyzed by highly specific proteases like TPA, the SK-plasminogen complex, and UK, all of which possess an unusually narrow protein substrate preference, namely a propensity to cleave the scissile peptide bond in PG. However, unlike UK and TPA, SK has no proteolytic activity of its own, and it activates PG to PN indirectly by first forming a high-affinity equimolar complex with PG, known as the activator complex (reviewed in Castellino, F. J., 1981, Chem. Rev. 81: 431).
Of the several thrombolytic agents used clinically, SK is probably one of the most-often employed, particularly because of its markedly lower cost when compared to TPA and UK. However, the choice of thrombolytic agent during therapy is dictated by a number of factors besides cost, such as the presence of side-effects and their severity, in vivo metabolic survival of the drug (e.g., circulatory clearance rates), fibrin selectivity and/or affinity, immunological reactivity etc. SK is a highly potent PG activator, and has a relatively long plasma half-life—properties that, together, impart a certain advantage to this drug as compared to its counterparts viz, TPA and UK. However, due to a lack of any appreciable fibrin clot-specificity in the former, the administration of therapeutically effective doses of SK often results in systemic PG activation, resulting in hemorrhagic complications due to the proteolytic degradation of blood factors by the plasmin generated throughout the circulatory system. However, if a fibrin affinity and/or selectivity could be integrated into SK a molecule which otherwise possesses little fibrin affinity of its own, it would considerably enhance the therapeutic efficacy of this thrombolytic agent. With respect to the other covered trait in a fibrinolytic agent, such as that described above for TPA above (viz., considerably lowered activity while circulating in the vascular system but enhanced PG activating ability in the presence of fibrin), attempts have been made in the past to produce analogs of SK with greater circulatory half-lives and decreased systemic plasmin generation by incorporating properties such as a slower rate of PG activation into the fibrinolytic agent. One example where this has been successfully accomplished is that of anisoylated streptokinase plasmin activator complex, abbreviated APSAC (sold under the trade-name ‘Eminase’ by the Beecham pharmaceutical group) (reference: Smith, R. A. G., Dupe, R. J., English, P. D., and Green, J., 1981, Nature 290:505) in which the catalytically important serine residue of the plasmin component is blocked by reversible acylation. The generalized plasmin activation coincident with the administration of unmodified SK has been reported to be appreciably diminished during thrombolytic therapy with APSAC since the deacylation of the covalently modified serine in the SK-acylated plasmin complex occurs slowly in the vascular system.
It is thus generally recognised that it will be of significant clinical advantage if SK could be engineered to possess increased fibrin affinity/specificity together with a markedly slower initial rate of activation of PG. Thus, soon after injection into the body, whilst it is still in an inactive or partially active state, such a modified PG activator will bind to the pathological fibrin clot during its initial sojourn through the vascular system in an inactive/partially active state. However, after an initial lag (a property engineered into the derivative/analog through design) it will become fully activated after being sequestered to the fibrin clot by virtue of its fibrin affinity. Thus, the PG activation process will be relatively limited to the immediate vicinity of the clot, thus obviating the systemic PG activation coincident with natural SK administration which has no intrinsic fibrin affinity of its own and which activates PG as soon as it encounters it. In other words, whilst the former property in the novel protein/s would be expected to confer on the thrombolytic agent an ability to target itself to the immediate locale of the pathological clot and thus help build up therapeutically effective concentrations of the activator therein, the initially slow kinetics of PG activation would result in an overall diminished generation of free plasmin in the circulation. The net result hall be a continued and more efficient fibrinolysis at the target sustained by considerably lowered therapeutically effective dosages of the thrombolytic agent.
In the past, the gene encoding for SK has been isolated from its natural source (Streptococcus species) and cloned into several heterologous micro-organisms such as yeast (Hagenson, M. J., Holden, K. A., Parker, K. A., Wood, P. J., Cruze. J. A., Fuke. M., Hopkins, T. R., Stroman, D. W., 1989, Enzyme. Microb. Technol. 11:650), bacteria viz, E. coli (Malke, H, Ferretti, J. J., 1984, Proc. Nat'l. Acad. Sci. 81: 3557), alternate species of Streptococcus (Malke, H., Gerlach, D., Kohler, W., Ferretti, J. J., 1984, Mol. Gen. Genet. 196:360), and Bacillus (Wong, S. L., Ye, R., Nathoo S., 1994, Applied and Env. Microbiol. 1:517). In addition, genetically modified SK derivatives containing “Kringle” type fibrin binding domains derived from plasminogen, and methods of obtaining the same by rDNA techniques, have been described (EU 0397 366 A1). However, since five such Kringl regions are already present in the natural SK-PG activator complex, being an integral part of PG in the activator complex, the advantages gained from further addition of such domains art likely to be minimal. Hence, there is a need to impart a qualitatively different fibrin-affinity and/or specificity to the activator complex, particularly of a type associated with TPA, a very effective thrombolytic agent possessing much greater fibrin affinity than SK. TPA is known to contain a fibrin-associating “finger” domain, which is structurally and functionally very similar to the fibrin-binding domains present in fibronectin, a multi-functional protein with ability to interact with a number of other proteins besides fibrin e.g., collagen, heparin, actin etc (reviewed in Ruoslahti. E., 1988, Ann. Rev. Biochem. 57:375). Methods for the imaging of fibrin-containing substances, such as pathological olots and/or atherosclerotic plaques in vivo by using large radio-labeled polypeptides derived from fibronectin, and gearing these FBDs (fibrin binding domains) have been disclosed (see: PCT WO 91/17765); this patent also discloses chemically cross-linked FBD-containing polypeptides and a thrombolytic agent (SK) to effect thrombus-targeted fibrinolysis. The chemical cross-linking procedure resulted in the generation of a complex mixture of heterogeneously cross-linked molecules with variable FED and SK content, since the bifunctional agents used for chemical cross-linking essentially cross-link any of the large number of lysine side-chains present in the participating molecules viz SK and HPG. Thus, this procedure generates mixtures of molecules with undefined location of the cross-links between the molecules e.g. both dimers and multimers with both hom—(e.g. SK-SK or FBD-FBD types) and hetero-crosslinked molecules with varying sites of cross-links are expected to be formed. In addition, it is noteworthy that the SK molecules chemically cross-linked with fibrin binding polypeptides disclosed in this patent showed an overall level of PG activator activity essentially comparable to that of unmodified SK, and no alteration was observed in the rate of PG activation, or the presence of an initial lag in the PG activation kinetics. It is quite clear that this invention related to the preparation of a heterogeneous population of cross-linked molecules with structures essentially undefined with respect to the cross-links' locations, and without any cross-correlation between the different structures in the ensemble of molecules and their corresponding functional properties. This is a serious limitation in the description of a drug intended for therapeutic application, in general, and with respect to the exact nature of the structure-function correlation in the collection of the cross-linked molecules, in particular.
In the past, hybrid SK derivatives with “kringle” type fibrin binding domains derived from human plasmin(ogen) fused to the former, and methods of obtaining the same by rDNA techniques, have been described (EU 0397 366 A1 and U.S. Pat. No. 5,187,098). However, five such Kringle regions are already present in the natural SK-Plasmin(ogen) activator complex, as noted before, being an integral part of PG in the activator complex, which has a weak fibrin affinity at best (Fears, R., 1989., Biochem. J. 261: 313). Hence, there is a need to impart a qualitatively different fibrin-affinity and/or specificity to the activator complex and utilize the affinity so imparted to obtain SK derivatives that display functional characteristics that help avoid the immediate activation of plasminogen upon contact with the latter.
Certain proteins are known to contain fibrin-associating “finger” domain/s, such as those present in fibronectin, a multi-functional protein with ability to interact with a number of other proteins besides fibrin e.g., collagen, heparin, actin etc (reviewed in Ruoslahti, E., 1988, Ann. Rev. Biochem. 57:375). TPA also possesses a “finger” type fibrin binding domain (FBD) that greatly helps in its fibrin association (Verheijen, J. H. et al., 1986., EMBO J. vol. 5, pp. 3525). Methods for the imaging of fibrin-containing substances, such as pathological clots and/or atherosclerotic plaques in vivo by using large radio-labeled polypeptides derived from fibronectin, and bearing these FBDs have been disclosed (see, PCT WO 91/17765); this patent also discloses chemically cross-linked FBD-containing polypeptides and a thrombolytic agent (SK) to effect thrombus-targeted fibrinolysis. However, it is noteworthy that the SK molecules chemically cross-linked with fibrin binding polypeptides showed an overall level of PG activator activity essentially comparable to that of unmodified KS, and no alteration was observed in the rate of PG activation or the presence of an initial lag in the PG activation kinetics. Besides, the cross-linking procedure resulted in the generation of a complex mixture of heterogeneously cross-linked molecules with variable FBD and SK content, since the bifunctional agents essentially cross-linked any of the large number of lysine side-chains present in the participating molecules viz. SK and HPG likely generating both dimers and multimers with both homo- (e.g., SK-SK or FBD-FBD types) and hetero-crosslinked molecules. Moreover, this invention essentially disclosed the preparation of a heterogeneous population of chimeric molecules between SK and fibrin binding polypeptide with undefined covalent structures with respect to the sites of cross-linking well as types of polymers so formed i.e. whether homo- (SK-SK or FBD-FBD types) or hetero-types, so that any meaningful structure-functional cross-correlation between the different structures in the ensemble and their corresponding functional properties cannot be obtained. This is a serious limitation in a drug intended for therapeutic application particularly one administered through a parenteral route in human beings.
In contrast the present invention provides novel clot-specific streptokinase proteins possessing altered plasminogen activation characteristics and a process for the preparation of different types of said proteins by recombinant DNA technology which have been designed using precisely defined elements of DNA polynucleotides that encode for fibrin binding domains and SK, or their modified forms. The hybrid proteins so formed thus have two very important structural as well as functional elements, namely SK or its modified forms, and finger type fibrin binding domain/s attached to each other through covalent peptide bonds in a predefined and predetermined order of juxtaposition with respect to each other (see FIG. 1 for types of such constructs, and the rationale for their construction, which is provided below) so that the hybrid, or chimeric, proteins so produced after expression in a suitable system possess discrete, definable covalent structures. In other words, the novel hybrid proteins contain SK or functionally relevant parts thereof, connected through polypeptide linkage/s with the relevant protein domains of human fibronectin that are capable of independently conferring fibrin affinity to the resultant hybrids in such a manner that the hybrid protein/s specifically display altered plasminogen activation characteristics. The latter is marked by the presence of an initial period of lag of several minutes' duration in the rate of PG activation by the hybrid SK derivatives (viz., time-delayed PG activation), which is followed by high rates of PG activation akin to that displayed by unmodified SK. In other words, the duration of the initial lag, which varies depending on the type of hybrid construct, is rapidly followed by PC activation rates closely similar to that of natural type SK. The simultaneous presence of the afore-mentioned two distinct biochemical properties in the same cot-dissolver protein molecule renders these hybrid streptokinases as very useful drugs for targeted, time-delayed clot lysis during thrombolytic therapy.
The biologically active form of Streptokinase (SK) is either the SK-plasminogen or SK-plasmin molecule/s, formed in the circulatory system by the association of SK with endogenous plasminogen soon after its administration in vivo. This complex is also known as the activator complex, a highly specific protease that activates substrate molecules of plasminogen to plasmin, which proteolytically digests fibrin and helps restore blood circulation in occluded vessels (Castellino, C. J., 1981., Chem. Rev. 81:431). Unlike free SK, which does not possess fibrin affinity, this complex already possesses substantial fibrin affinity of its own due to the “kringle” fibrin binding domains present in the plasmin(ogen) part of the SK-plasmin(ogen) activator complex Fears R., 1989., Biochem. J. 261:313 see also references cited therein). Nevertheless, unlike other preferred plasminogen activator protein drugs such as tissue plasminogen activator (TPA) which possesses intrinsic fibrin affinity as well as a fibrin-dependent plasminogen activation kinetics, the administration of SK during clot dissolution therapy often leads to unwanted systemic activation of plasminogen throughout the circulatory system due to immediate activation of circulating plasminogen, as opposed to the desired activation in and around the fibrin clot occluding the flow of blood through the affected vessel/s.
Thus, it will be of significant clinical advantage if SK could be engineered to possess increased fibrin affinity/specificity together with a markedly slower initial rate of activation of plasminogen (PG), but becoming capable of activating plasminogen in a manner similar to that of unmodified SK after an initial hiatus. Thus, soon after injection into the body, whilst it is still in an inactive or partially active state, the engineered SK will bind to the pathological fibrin clot while still in an inactive or partially active state, as it sojourns through the vascular system by virtue of the engineered fibrin affinity. However, after the initial lag in its PG activation kinetics is overcome in a few minutes, it will preferentially become activated in the immediate vicinity of the clot where it is now sequestered, thereby obviating or significantly minimizing the systemic PG activation coincident with natural SK administration which immediately activates PG upon administration. Thus, whilst the former property (of fibrin affinity) would be expected to confer on the new thrombolytic agent an ability to target itself to the immediate locale of the pathological clot and thus help build up therapeutically effective concentrations of the activator therein, the other property (of an initially slow kinetics of PG activation) would result in an overall diminished generation of free plasmin in the circulation. The net result shall be a continued and more efficient fibrinolysis at the target sustained by considerably lowered therapeutically effective dosages of the thrombolytic agent. In conclusion, a fibrin affinity per se in SK has little beneficial consequences (which anyway the SK-PG complex possesses in some measure) unless the systemic PG activation is thwarted and/or delayed.
An important attribute of the present invention is the preparation of different types of novel and hitherto undisclosed chimeric SK derivatives produced by recombinant DNA technology using defined gene-segments of SK and FBD combined in a pre-designed manner. These novel genetic constructs have been designed using precisely defined DNA elements that encode for SK and fibrin binding domains, or their modified forms so as to retain the functional characteristic of each (PG activation and fibrin affinity, respectively) as well as a characteristically altered PG activation kinetics. The chimeric proteins so produced have two types of elements (SK and the ‘finger’-type fibrin binding domains, or their modified forms) in a predefined and predetermined order of juxtaposition with respect to each other, so that the chimeric proteins expressed from these genes possess discrete, definable covalent structures. In other words, the chimeric proteins contain SK or parts thereof, connected through polypeptide linkage with the relevant protein domains that confer fibrin affinity to the resultant hybrids and also specifically result in altered kinetics of PG activation. The latter is characterized by an initial lag, or absence of PG activation, of several minutes' duration (viz., time-delayed PG activation), followed by high rates of PG activation akin to that of unmodified SK. The initial lag (which varies from approx. 8 min to 25 min depending on the design of the SK derivative) is rapidly followed by high rates of PG activation closely similar to that displayed by natural type SK. The simultaneous presence of the afore-mentioned two biochemical properties in the same PG activator molecule has not been disclosed in the SK-derived molecules in either of the patent disclosures cited above. In addition, the present patent discloses new combinations of DNA sequences that have been used to express the novel protein molecules with a unique combination of functional properties, mentioned above, which are not disclosed in the other patents.
The rationale for the construction of hybrid SK derivatives as disclosed in the process of the present invention with both fibrin specificity and delayed PG activation kinetics is explained below.
The molecular basis for the fibrin affinity displayed by fibronectin has been studied in some detail in recent years (Matsuka, Y. V., Medved, L. V., Brew. S. A. and Ingham; K. C., 1994, J. Biol. Chem. 269:9539). Under physiological conditions, FN first interacts reversibly (but with relatively high affinity) with fibrin and is then covalently incorporated into the fibrin clot matrix through clotting factor XIII, a transglutaminase (reviewed in Ruoslahti, E., 1988, Anna. Rev. Biochem. 57:375), whose action results in the covalent cross-linking between FN and a lys residue in fibrin(ogen) at the reactive Gln (residue 3) of the former. The region/s responsible for the interaction of FN with fibrin have been identified to resin both in the N-terminal as well as the C-terminal ends of this multi-domain protein. The N-terminal region of FN comprises of five finger modules (FBDs) as well as a transglutaminase cross-linking (TG) site, whereas the C-terminal region, lacking a TG site, contains three modules, as demonstrated by the binding of different polypeptides derived from FN carrying these two broad regions to fibrin-agarose. The exact domains in the N-terminal region responsible for the strong binding of the FN module, and their relative contributions towards this interaction have been analysed closely (Matsuka, Y. V., Medved, L. V., Brew, S. A. and Ingham, K. C., 1994, J. Biol. Chem. 269:9539 and Rostagno et al., 1994; J. Biol. Chem. 269, 31938) by expressing DNA segments encoding various combinations of the modules in heterologous cells and/or by examining the fibrin binding properties of polypeptide fragments carrying these modules prepared by limited proteolysis of FN. These studies clearly identified that of all the individual modules present in the N-terminal region of FN, the bi-modular arrangement viz., FBD 4 and 5 domains, displayed a fibrin affinity significantly comparable to the interaction of the full-length FN molecule, in contrast to all the other domains either as pairs or individually (including 4 and 5) which displayed poor affinity at 37° C. It is therefore clear from these studies that physiologically effective fibrin binding is not a common property of all the modules, either individually or in pairs, but is principally located in the FBD pair of 4 and 5, and to a relatively lesser extent, in domains 1 and 2.
To achieve the functional objective of an initially time-delayed PG activation kinetics by the hybrid SK derivatives, our design utilizes the fusion of selected regions of the FBDs of human fibronectin or its homologous sequences present in other proteins with SK (or its partially truncated forms) at strategically useful points so as to kinetically hinder the initial interaction of SK with PG necessary to form the 1:1 stoichiometric activator complex. It is known that of the 414 residues constituting native SK, only the first 15 residues and the last 31 residues are expendable, with the resultant truncated polypeptide being nearly as active as the native full-length protein in terms of PG activation ability. Further truncation at either end results in drastic decrease in the activity associated with the molecule (Malke, H., Roe, B., and Ferretti, J. J. (1987) In.: Streptococcal Generics, Ferretti, J. J., and Curtis, R. III [Ed.] Proc. American Society for Microbiology., Washy D. C. p. 143). It has been demonstrated that SK interacts with PG through a least two major loci, mapped between residues 16-51 and 230-290 (Nihalani, D., Raghava, G. P. S., and Sahni, G., 1997, Prot. Sci. 6:1284), and probably also the region and around residues 331-332 (Lin, F. L., Oeun, S., Houng, A., and Reed, G. L., 1996, Biochemistry 35:16879); in addition the sequences at the C-terminal ends, especially before the last 30-32 residues of the native sequence (Kim. J. C., Kim, J. S., Lee, S. H., and Byun, S. M., 1996, Biochem. Mol. Bio. Int. 40:939. Lee, S. H., Jeong, S. T., Kim, I. C., and Byun, S. M., 997, Biochem. Mol. Bio. Int. 41:199. Fay, W. P., Bokka, L. V., 1998, Thromb. Haemost 79:.985) are important in generating the activator activity associated with the complex. Since a primary consideration in designing the SK-FBD chimeras was engineering of a decreased, or kinetically slowed, initial PG activation rate, we reasoned that either the C- or N-termini (or both, together) could be utilized to bear the FBDs in the hybrid structures, and that the presence of such ‘extra’ domains in SK, either full-length or already truncated to the most functionally relevant regions of human fibronectin that can independently bind fibrin under physiological conditions (detailed earlier) and would also suitably retard and/or delay the PG activation rates observed normally with native SK and PG by interfering in the interactions of SK with PG to generate a functional activator complex. If the polypeptide in between these two distinct parts constituting the chimera were sufficiently flexible, proteolytic scission in this region would then result in the removal of the retarding/inhibiting portion (FBD component) from the SK-FBD hybrid and lead to a burst of PG activation after an initial delay. This proteolysis could be mediated by the small amounts of endogenous plasmin generated in the vicinity of the pathological clot by intrinsic plasminogen activator/s of the system, such as TPA, urokinase etc already present in the circulatory system. Indeed, this premise was borne out by experimentation, which showed that the lag times of PG activation by the SK-FBD chimeras disclosed in this invention were directly governed by plasmin-mediated proteolysis of the hybrid proteins leading to the liberation of the FBD portion from the SK-FBD followed by rapid PG activation by the SK. The direct implication of this functional property in a plasminogen activator is that once injected into the body, the protein could then traverse in an inactive state through the circulatory system and bind to the pathological clot by virtue of the fibrin affinity imparted by the fibrin binding domains thereby obviating or minimizing systemic PG activation. Thus, if the thrombolytic agent traverses the circulation prior to this activation (which is known to require 3-5 minutes in the human circulation), the fibrin affinity in the chimera would allow it to bind to the clot, thereby helping to localize the PG activation in and around the immediate vicinity of the thrombus.
The amino acid sequence of human FN is known to be composed of three types of homologous repeats (termed type-1, type-2 and type-3), of which the FBDs at the amino terminus of FN are made of five type-1 repeats, each approximately 50 residues long and containing two disulfide bridges. The C-terminus of FN also has three type-2 homology repeats that are involved in fibrin-FN interactions. Therefore, altogether, a large portion of the FN molecule, representing the several N- and C-terminally located FBDs, could be linked with SK if all of the fibrin interacting regions need to be incorporated into the contemplated SK-FBD chimeras. However, such a design produces a chimeric protein that is not only too bulky, but also decreases the probability for the polypeptide to fold into a biologically active conformation due to the presence of a large number of S—S bridges that may form non-native, intra- and inter-molecular dude bonds. Instead, a potentially more worthwhile proposition is to seek miniaturised but, nevertheless, functionally active combinations of selectively truncated regions of SK and/or FBDs.